cDNA amplification from microscopic amounts on animal tissue
    From: Generation of cDNA libraries: Methods and protocols (Shao-Yao Ying Ed.) Humana press, pp. 103-116.Methods_files/Matz%20cDNA%20amplification%202003.pdfMethods_files/Porites%20Stress%20DB-3.htmlshapeimage_2_link_0
    From: Generation of cDNA libraries: Methods and protocols (Shao-Yao Ying Ed.) Humana press, pp. 41-50.
Analysis of OTU abundances using Bayesian GLM models
   Download the MCMC.OTU package from CRAN by saying in R:
    Please report bugs to matz[at]!
  The package uses MCMC.qpcr philosophy to analyze up to about 200 OTU abundances in one- or two-way designs based on counts.
   Walkthrough with explanations and examples:
    Reference:  Green et al PeerJ 2014  
Bayesian analysis of qRT-PCR data: no control genes required
Install the latest version of the MCMC.qpcr R package from CRAN by saying in R:
mcmc.qpcr.tutorial.pdf : A tutorial for the MCMC.qpcr package.
2b-RAD: cost-efficient whole-genome genotyping
   Original 2bRAD paper:  wang2012_NatureMethods_2b-RAD.pdf
       2bRAD_protocol.pdf   (version July 2014) 
        Bioinformatics (version September 2014):
            de novo
            with reference genomeMethods_files/wang12%202b-RAD.pdfMethods_files/2bRAD_protocol-1.pdf
tag-based RNA-seq: global gene expression profiling for <$50 / sample 
(now using Illumina HiSeq)
  scripts, instructions and lab protocol 
    pdf of the paper: Meyer et al Mol Ecol 2011
Gene Ontology analysis with Mann-Whitney U tests
and adaptive clustering
Scripts and example data  (version July 2014)
On a Mac or Linux/Unix, a single script will run it all: GO_MWU.R (but make sure all other scripts are also saved in the same working directory)  
We used this approach in the paper by Voolstra et al, PLoS ONE 2011, 6(5): e20392 ; see references therein for the earlier applications. The method is applicable to any analysis involving ranked lists of GO-annotated genes, such as rates of evolution or differential expression. Its advantage over the more conventional  “GO enrichment analysis” is twofold: (1) it makes use of the whole dataset rather than only of the genes passing an arbitrary significance cutoff; (2) it simplifies and adapts the GO hierarchy to the particular dataset by grouping and clustering GO categories based on sharing of annotated genes.  Basic knowledge of R and manual editing of the resulting images is required, however.  
  As of July 2013, the method accepts binary measures, i.e. 1 or 0 depending on whether the gene is a member of some group, such as a co-expression module, or not. In this case Fisher’s exact test is performed instead of MWU test; the clustering is still done.
Sequencing and de novo transcriptome analysis using 454
Contact matzlab for HiSeq-ADAPTED protocol
   cDNA preparation protocol  
   (for 454 Titanium)  
   bioinformatic scripts
   step-by-step bioinformatics guide
   BMC genomics paper
Metabarcoding based on sequenced amplicons 
   This simple pipeline is based on cd-hit-otu and was originally designed to process Symbiodinium ITS amplicon data, but it can be used for any amplicons. For statistical analysis of counts derived by this pipeline, see MCMC.OTU package below.Methods_files/MetaBarCDHIT.tgz
Assembly of a transcriptome de novo using Trinity and its annotation
The annotaton pipeline involves blastx to a local copy of a uniprot (uniref50) database to assign gene names and GO annotations and to extract CDS (protein-coding sequences). The output of the CDS_extractor script is used to assess the completeness of the transcripts (distribution of assembled CDS length compared to the length of reference proteins). KOG and KEGG annotations are assigned using web-based tools (see included walkthrough).
KOG-MWU analysis: summarize the whole RNA-seq experiment in 23 lines 
   KOGMWU is now an R package on CRAN  
   This method is very similar to GO-MWU analysis above, except it operates with KOG class annotations and produces a simple 23-line table (there are only 23 meaningful KOG classes). The goal is to determine which KOG classes are significantly enriched by differentially expressed genes. The delta-ranks reported in the table can be used to compare overall pictures of gene expression across diverse datasets. See examples and comments within the R script.
Transcriptome contiguity analysis
Based on the transcritome’s blastx hits to a reference protein database, the goal is to plot the distribution of reference coverages by the longest matching contig and calculate contiguity as per Martin and Wang, Nat Rev Genet  2011 (proportion of contigs with coverage exceeding certain threshold)